Methods and apparatus for image processing of spherical content via hardware acceleration components. In one embodiment, an EAC image is subdivided into facets via existing software addressing and written into the memory buffers (normally used for rectilinear cubemaps) in a graphics processing unit (GPU). The EAC facets may be translated, rotated, and/or mirrored so as to align with the expected three-dimensional (3D) coordinate space. The GPU may use existing hardware accelerator logic, parallelization, and/or addressing logic to greatly improve 3D image processing effects (such as a multi-band blend using Gaussian blurs.)
Legal claims defining the scope of protection, as filed with the USPTO.
3. The computerized apparatus of claim 1, where the graphics processing unit comprises wrapping logic that treats certain pixels of the plurality of buffers as neighbor pixels.
4. The computerized apparatus of claim 3, where the graphics processing unit comprises linear blurring logic that can process the neighbor pixels without checking for wrapping.
5. The computerized apparatus of claim 3, where the graphics processing unit comprises multi-band blurring logic that can handle non-power-of-two remainders.
6. The computerized apparatus of claim 3, where the first camera determines a first exposure setting for the first image and the second camera determines a second exposure setting for the second image.
7. The computerized apparatus of claim 6, where the one or more hardware accelerated image processes comprises blurring an exposure transition across the equi-angular facets.
8. The computerized apparatus of claim 7, where blurring the exposure transition approximates a Gaussian blur with a plurality of spatially weighted linear blurs.
10. The method of claim 9, further comprising writing the processed equi-angular facets to a processed equi-angular cubemap data structure and encoding the processed equi-angular cubemap data structure for transfer to another device.
11. The method of claim 9, further comprising determining a first exposure setting for the first image and determining a second exposure setting for the second image.
12. The method of claim 11, where the one or more hardware accelerated image processes comprises blurring an exposure transition across the equi-angular facets.
13. The method of claim 12, where blurring the exposure transition approximates a Gaussian blur with a plurality of spatially weighted linear blurs.
14. The method of claim 13, where the one or more hardware accelerated image processes can handle non-power-of-two remainders.
15. The method of claim 11, where the one or more hardware accelerated image processes treat certain pixels of the corresponding memory structures as neighbor pixels.
17. The non-transitory computer-readable apparatus of claim 16, where the plurality of instructions further cause the computerized apparatus to perform at least one image transformation.
18. The non-transitory computer-readable apparatus of claim 17, where the at least one image transformation comprises three image rotations.
19. The non-transitory computer-readable apparatus of claim 16, where the plurality of instructions are further configured to write the processed equi-angular facets to a processed equi-angular cubemap data structure and encode the processed equi-angular cubemap data structure for transfer to another device.
20. The non-transitory computer-readable apparatus of claim 19, where the equi-angular cubemap data structure comprises image data that is oriented to minimize compression artifacts.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
September 8, 2022
January 30, 2024
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.